fais

PaperMind AI is a modern PDF document chat application powered by AI that allows you to upload PDF documents and have intelligent conversations about their content..

OpenAI-compatible LLM inference gateway with exact (Redis) and semantic (FAISS) caching and calibrated hit probabilities. Phase 2 of a thesis on cost-quality-latency routing.

RAG-powered Amazon Beauty product search and recommendation system using BM25, semantic search, hybrid retrieval, and Llama 3.3 70B via Groq API.

is a Streamlit-powered web app that lets you upload any PDF and ask natural-language questions about its contents

Fais is a lightweight Node.js web framework designed for building server-side applications. It provides a simple and intuitive interface for defining routes, handling HTTP requests, and serving static assets. With support for middleware and request body parsing, Fais enables developers to create robust and scalable web applications with ease.

This repo contains most of the work I have done during my summer internship 2018 at FAIS, HGC, University of Tokyo.

A local RAG-based research paper assistant using Streamlit, LangChain, FAISS, and Ollama. Upload PDFs and ask context-aware questions with local LLMs.

AI-powered relationship CRM built as 4 Spring Boot microservices with Kafka-async AI enrichment. AWS Bedrock + FAISS semantic dedup improves contact-matching accuracy by 70%; a transparent 0-100 relationship-health score updates in under 200ms, and Claude (200K-token context) answers natural-language CRM queries in 2-5 seconds.